Documentation for a newer release is available. View Latest

HTM Server Deployment Configuration

1. Image Name and Version

To deploy the Human Task Manager (HTM) server, you need to pull the Docker image from the ICON Solutions container registry.

Image name:

registry.ipf.iconsolutions.com/human-task-manager-app

Versioning:

We follow semantic versioning using the MAJOR.MINOR.PATCH format (e.g. 1.3.9). You should always reference a specific version tag to ensure deployment consistency and avoid unintended updates.

Example configuration

image:
  name: registry.ipf.iconsolutions.com/human-task-manager-app
  tag: 1.3.9

Avoid using the latest tag in production environments.

2. Example Kubernetes Manifests

The following example demonstrates a basic deployment of the Human Task Manager (HTM) server using Kubernetes manifests. These can be used as a starting point for writing Helm charts or direct K8s deployments.

This example includes:

ConfigMap for application configuration
Deployment for the HTM application
Service for exposing the HTM app internally
Ingress configuration

ConfigMap: htm-cm

apiVersion: v1
kind: ConfigMap
metadata:
  name: htm-cm
data:
  application.conf: |-
    ipf.mongodb.url = "${ipf.mongodb.url}"
    ipf.htm {
      event-processor.delegating.enabled = ${ipf.htm.event-processor.delegating.enabled}
      mongodb.purging.time-to-live = 15minutes
    }

    management.health.mongo.enabled = false

    event-processor {
      restart-settings {
        min-backoff = 500 millis
        max-backoff = 1 seconds
      }
    }

    akka {
      actor.provider = cluster

      remote.artery.canonical.hostname = ${POD_IP}
      # Use Kubernetes API to discover the cluster
      discovery {
        kubernetes-api {
          pod-label-selector = "app=%s"
        }
      }

      management {
        # use the Kubernetes API to create the cluster
        cluster.bootstrap {
          contact-point-discovery {
            discovery-method          = kubernetes-api
            service-name              = ${AKKA_CLUSTER_BOOTSTRAP_SERVICE_NAME}
            required-contact-point-nr = 1
            required-contact-point-nr = ${?REQUIRED_CONTACT_POINT_NR}
          }
        }
        # available from Akka management >= 1.0.0
        health-checks {
          readiness-path  = "health/ready"
          liveness-path   = "health/alive"
        }
      }

      cluster {
        seed-nodes = []
        downing-provider-class = "akka.cluster.sbr.SplitBrainResolverProvider"
        split-brain-resolver {
          active-strategy = keep-majority
          stable-after = 20s
        }
        sharding {
          remember-entities = off
          handoff-timeout = 8s
          least-shard-allocation-strategy.rebalance-absolute-limit = 20
          rebalance-interval = 2s
          number-of-shards = 100,
          distributed-data.majority-min-cap = ${akka.management.cluster.bootstrap.contact-point-discovery.required-contact-point-nr}
          snapshot-after = 50
        }
      }
    }

  logback.xml: |-
    <?xml version="1.0" encoding="UTF-8"?>
    <configuration>
        <appender name="CONSOLE" class="ch.qos.logback.core.ConsoleAppender">
            <target>System.out</target>
            <encoder>
                <pattern>[%date{ISO8601}] [%level] [%logger] [%marker] [%thread] - %msg%n</pattern>
            </encoder>
        </appender>

        <logger name="akka.stream.scaladsl.RestartWithBackoffSource" level="warn"/>
        <logger name="com.iconsolutions.ipf.core.connector" level="INFO"/>
        <logger name="com.iconsolutions.ipf.core.platform.read.processor" level="INFO"/>
        <logger name="com.iconsolutions.akka.persistence.mongodb" level="INFO"/>
        <logger name="com.iconsolutions.ipf.htm.notification.autoconfig.TaskNotificationAutoConfiguration" level="WARN"/>
        <logger name="com.iconsolutions.ipf.htm.task.task_manager.handlers.commands.processors.builtin.TaskManagerActionRevivalProcessor" level="DEBUG"/>

        <appender name="ASYNC" class="ch.qos.logback.classic.AsyncAppender">
            <queueSize>8192</queueSize>
            <neverBlock>true</neverBlock>
            <appender-ref ref="CONSOLE" />
        </appender>

        <root level="info">
            <appender-ref ref="ASYNC"/>
        </root>
    </configuration>

Service and Deployment

apiVersion: v1
kind: Service
metadata:
  labels:
    app: human-task-manager
    product: ipfv2
  name: human-task-manager
spec:
  ports:
    - name: server-port
      port: 8080
      protocol: TCP
      targetPort: 8080
  selector:
    app: human-task-manager
    product: ipfv2
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: human-task-manager
  annotations:
    prometheus.io/scrape: "true"
    prometheus.io/path: "/"
    prometheus.io/port: "9001"
  labels:
    app: human-task-manager
    product: ipfv2
spec:
  replicas: 1
  selector:
    matchLabels:
      app: human-task-manager
      product: ipfv2
  template:
    metadata:
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/path: "/"
        prometheus.io/port: "9001"
      labels:
        app: human-task-manager
        product: ipfv2
      name: human-task-manager
    spec:
      containers:
        - name: htm
          image: registry.ipf.iconsolutions.com/human-task-manager-app
          imagePullPolicy: IfNotPresent
          ports:
            - name: http
              containerPort: 8080
            - name: debug-port
              containerPort: 5005
          livenessProbe:
            failureThreshold: 5
            httpGet:
              path: /actuator/health/liveness
              port: http
              scheme: HTTP
            initialDelaySeconds: 30
            periodSeconds: 2
            successThreshold: 1
            timeoutSeconds: 1
          readinessProbe:
            failureThreshold: 5
            httpGet:
              path: /actuator/health/readiness
              port: http
              scheme: HTTP
            initialDelaySeconds: 30
            periodSeconds: 10
            successThreshold: 1
            timeoutSeconds: 2
          env:
            - name: "POD_NAME"
              valueFrom:
                fieldRef:
                  fieldPath: metadata.name
            - name: "IPF_JAVA_ARGS"
              value: " -XX:+UseContainerSupport -XX:MaxRAMPercentage=60 -XX:InitialRAMPercentage=60
            -XX:-PreferContainerQuotaForCPUCount -Dma.glasnost.orika.writeClassFiles=false
            -Dma.glasnost.orika.writeSourceFiles=false"
            - name: "POD_IP"
              valueFrom:
                fieldRef:
                  fieldPath: status.podIP
            - name: "KUBERNETES_NAMESPACE"
              valueFrom:
                fieldRef:
                  fieldPath: metadata.namespace
            - name: "AKKA_CLUSTER_BOOTSTRAP_SERVICE_NAME"
              valueFrom:
                fieldRef:
                  fieldPath: metadata.labels['app']
          resources:
            requests:
              memory: 4Gi
              cpu: "4"
            limits:
              memory: 4Gi
          volumeMounts:
            - mountPath: /human-task-manager-app/conf/logback.xml
              name: config-volume
              subPath: logback.xml
            - mountPath: /human-task-manager-app/conf/application.conf
              name: config-volume
              subPath: application.conf
      volumes:
        - name: config-volume
          configMap:
            name: htm-cm
        - name: keystore
          secret:
            secretName: keystore

Ingress Example

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: human-task-manager
  annotations:
    cert-manager.io/cluster-issuer: letsencrypt
    kubernetes.io/ingress.class: nginx
spec:
  rules:
    - host: htm.${ipf.ingress-domain}
      http:
        paths:
          - path: /
            pathType: Prefix
            backend:
              service:
                name: human-task-manager
                port:
                  number: 8080

Placeholders such as ${ipf.mongodb.url} and ${ipf.ingress-domain} must be replaced with actual values specific to your environment before deployment.

These are minimal manifests for demonstration. You may need to extend them based on your infrastructure (e.g. secrets, TLS, persistent storage).

3. Infrastructure Requirements

The HTM server depends on external infrastructure components that must be available and configured correctly prior to deployment.

MongoDB

HTM uses MongoDB as its primary data store.

You must provision a MongoDB instance accessible from the cluster.
The MongoDB connection string should be provided via the configuration file using the following placeholder ipf.mongodb.url

Kafka

HTM can send notifications when tasks are closed, using Apache Kafka.

You must have a Kafka cluster accessible from your environment.
Create Kafka topics for closed task notifications and task registrations.
Configure the topic names in the config file via variable htm.kafka.producer.topic and ipf.htm.async.register-task.kafka.producer.topic

4. Base Resource Requirements

To ensure stable performance, the HTM server requires the following baseline resources:

CPU: Minimum of 4 cores
Memory: Recommended 4 GB

These values should be considered the lower bound for production environments, the actual values should always be determined by running adequate game-day load tests.